Case Study
Novo Quote is Novo's auto insurance quoting flow — users answer questions, get a rate, and sign up for coverage.
The original design was long forms without pre-fills or smart logic to help user skip questions. Most users dropped without getting a quote.
Redesigned the flow question-by-question, added data pre-fill, removed the Submit button, and introduced conditional logic to skip irrelevant steps.
Overview
Novo Quote is Novo's primary customer acquisition product — a web-based quoting flow where prospective drivers answer a series of questions, get an auto insurance quote, and sign up for coverage. It's the main way Novo brings in new policyholders.
Users arrive from multiple channels: lead partners like Insurify and MediaAlpha, paid ads on Google and Meta, and organic traffic from the Novo website.
When I joined in late 2024, the flow was underperforming. Many users were dropping off before finishing. The form felt long, some fields were confusing, and there were bugs to fix. There are significant opportunities to improve.
The Problem
"84% of insurance leads abandon their quotes — the highest abandonment rate of any industry, exceeding even e-commerce. The leading causes: multi-page forms without progress indicators, too much information required upfront, and poor overall digital experience."
ProPair / Industry Research
The original experience put all the questions on one long scrolling page — a dense form that asked users to process and fill out everything at once before moving on. Most didn't make it to the end.
There was also no data pre-fill. Even when Novo already had information about a user — from a lead partner or a background check — they still have to type manually. That friction alone was a significant driver of abandonment. On top of that, every user saw every question regardless of their situation, there was no way to skip ahead, and users had no visibility into how far along they were or whether they'd even qualify.
What made this especially challenging was the underlying complexity of the problem space. Insurance quoting comes with strict regulatory and business requirements around which questions must be asked, and in what order. And depending on how a user answers, the flow can branch into a very different path — different follow-up questions, different eligibility outcomes, different product options. That branching compounds quickly. Designing an experience that handles all of that complexity while still feeling simple and linear is central to the problem we want to address.
| Before | After | |
|---|---|---|
| All questions on one long scrolling page | → | One question at a time, focused flow |
| No data pre-fill — users entered everything manually | → | Data pre-fill from lead partners and background checks |
| Static form with no conditional logic | → | Smart conditional logic skips irrelevant steps |
| Users had to tap through every field | → | Tap to select = instant progression, no Submit button |
| Rate shown only at the very end | → | Rate estimate shown early, updates at key moments |
| Ineligible users discovered too late | → | Ineligible users flagged early in the funnel |
| Confusing copy and interface bugs | → | Simplified, user-friendly copy throughout |
| No sense of progress or momentum | → | Progress bar creates constant sense of momentum |
Design Process
A question-by-question redesign meant rebuilding the entire funnel from scratch — not a small ask. Before that could happen, I needed evidence to make the case.
I started with competitive analysis, looking at how other insurance and fintech products handle multi-step flows. I pulled Novo's product analytics to map where users were dropping. But the most convincing research came from session replays. Watching real users navigate the form made the problem visceral: people weren't just leaving because the flow was long. They were manually entering information Novo already had.
Armed with that evidence, I made the case to the PM and stakeholders, grounding the pitch in session data, drop-off rates, and competitive precedent. On top of that, I also tackled the smaller usability debt — unclear labels, broken interactions, confusing copy. Those fixes mattered just as much to the final completion rate as the bigger redesign.
Rather than a full rebuild at once, we phased the rollout over three months, which let us validate early and keep learning as we went. Here's how the process looked like.
Competitive analysis, product analytics, session replay review, and getting stakeholder buy-in on the question-by-question direction and pre-fill strategy.
Designing the conditional branching logic, defining which questions could be skipped, and establishing the entry experience for different traffic sources.
Rolling out the redesigned flow in stages across the funnel over three months, iterating based on performance data along the way.
Refining copy, improving rate estimate placement, flagging ineligible users earlier, and polishing the overall experience.
The Solution
A long scrolling form asks users to take in the entire problem at once. A question-by-question flow asks them to answer one thing, then move on. The cognitive difference is significant — and it shows in completion rates.
We also removed the "Next" button entirely. Tapping an answer automatically advances the flow; a "Back" button is always available. That single interaction change cut the number of required taps roughly in half.
"Simplifying registration forms consistently produces hugely increased conversion rates. Removing friction from any step of a form reliably improves completion — and form usability improvements are one of the strongest ROI arguments for UX investment."
Nielsen Norman Group
Every answer advances the bar and transitions the screen. It sounds minor. In practice, it's one of the stronger psychological levers in the design — users feel momentum with every tap rather than weighing whether it's worth continuing.
Answers trigger conditional branching behind the scenes. If a question isn't relevant, users skip it without knowing. If more detail is needed, a follow-up appears. The flow adapts to each user's situation while still feeling linear and simple.
We surfaced the rate estimate earlier so users aren't waiting until the final screen to see a number. Rather than updating it in real time — which risks showing the price climb with every driver or vehicle added — we display it at specific moments. Informed without anxious.
Users arriving from lead partners may already have information on file. A background check at entry lets us prepopulate fields and skip questions they don't need to answer. Everyone only sees what's actually relevant to their situation.
Session replays made it clear: asking users to manually re-enter data Novo already had was one of the single biggest drivers of abandonment. The redesign addressed this directly — wherever data is available, it's pre-filled and surfaced for confirmation rather than re-entry. Users only type what we genuinely don't know. The impact on completion rates was outsized, especially in the vehicle and driver sections.
Previously, some users completed the entire flow only to find out at the end they didn't qualify. We now identify and notify them earlier — saving their time and improving the quality of applications that make it through.
The Flow
Eight steps from first landing to active policyholder. Each one designed to ask only what's necessary, pre-fill what we already know, and keep the user moving.
The user lands and sees what we already know about them — or a clean slate if they're coming in cold. A background check runs in the background to assess risk and prepopulate data where possible.
All household members and potential drivers are added to the policy. Known information surfaces for confirmation rather than re-entry.
Vehicles to be covered are added. Each addition adjusts the rate estimate at the right moment.
Driving history is collected or pulled from the background check. Occupation, prior coverage, and lapse of insurance round out the risk profile.
Users choose from three product variants. The Novo Safety Program is introduced here — a telematics approach that uses Bluetooth to track driving behaviour and adjust pricing over time.
The user reviews their quote in full. Changes can still be made. Last checkpoint before commitment.
Electronic signature, email verification, payment. The user is now a Novo policyholder.
Users download the Novo app and complete vehicle setup to activate the telematics safety program.
Results
Every key metric improved from Q1 to Q2 2025.
The table compares Jan–Mar against Apr–Jun, with FY26 H1 targets for context.
| Jan – Mar 25 | Apr – Jun 25 | FY26 H1 Target | ||
|---|---|---|---|---|
| Quote Completion % | 16% | 50% | 3× | — |
| Quote to Bind % | 4.5% | 9% | 2× | — |
| Visitor to Bind % | 0.3% | 2.8% | 10× | 6% |
| # of Binds | 31 | 958 | 30× | +2,000 |
Beyond the headline numbers: time to quote dropped to around five minutes on average — with one user completing the full flow in under forty seconds. Customers who signed up also showed stronger risk profiles than before, a direct result of the smarter flow collecting better quality data.
Novo Quote is live and growing. Over 2,500 paying customers since launch — and counting. These are monthly subscribers, not free sign-ups. Each one represents a genuine commitment to Novo as their auto insurer.
A note on growth pace. Shortly after launch, the team noticed churn — customers signing up but not sticking around. In response, Novo pulled back on ad spend and tightened acquisition criteria, deliberately trading volume for quality. The 2,500 figure reflects that choice. It's a smaller number than aggressive growth would have produced, but these are customers worth keeping.
Reflection
The question-by-question pattern was the biggest lever — but it wasn't an obvious call. It required research to back it up, data to make the case, and stakeholder trust to execute. Releasing in phases was the right move: it let us prove the concept on a slice of the flow before committing to a full rebuild.
Some of the biggest wins came from small decisions. Removing the Submit button cuts at least one tap per question — across dozens of questions, that adds up to a meaningfully faster experience. It sounds trivial in isolation. It isn't.
Choosing not to update the rate in real time was equally deliberate. Showing a price that climbs with every driver added is technically accurate but emotionally corrosive. Surfacing it at the right moments — rather than constantly — made the flow feel considered rather than transactional.
What I didn't see coming was the churn. After a strong launch quarter, a meaningful number of customers dropped off within the first few months. Optimising hard for conversion had a side effect: we were also letting through users who weren't the right fit for Novo's product. The team responded by tightening acquisition criteria and pulling back on ad spend — trading volume for quality. It was the right call. It also sharpened my thinking about what good conversion design actually means: not just getting people through the funnel, but getting the right people through it.